[Company] Weekly Operations Update 3 and 4/2010

Polossatik polossatik at gmail.com
Tue Jan 26 05:37:23 EST 2010


On Tue, Jan 26, 2010 at 6:48 AM, Bas Wijnen <wijnen at debian.org> wrote:
> On Mon, Jan 25, 2010 at 08:29:58PM -0500, Wolfgang Spraul wrote:
>> ---2 unique IDs and serial numbers
>> I have a question: How do we, as the Qi Hardware project, feel about unique
>> numbers and IDs in and on our devices, in general?
>
> I think it may be useful to have a globally unique number printed on
> each unit, so it is possible to hold them apart.  I am opposed to having
> it in hardware, in a way that is readable by software (like the CPU ID
> feature of Intel CPUs).  I don't see any reasonable use for this (that
> can't be easily achieved without it).  It does make it hard to be
> anonymous, possibly needing kernel hacks to avoid it to be read out.
>
>> As a proposal, I would say we try to keep our products as anonymous as
>> possible.  Use unique IDs only when they provide real value, and
>> _ALWAYS_ disclose what kind of numbers we have in which place.
>
> Good idea.  IMO it's fine to not have it printed on the side as well,
> but having it there shouldn't be a problem: people know when it's read,
> and can remove the sticker, should they want to.
>
> Thanks,
> Bas

I don't see the need for real hardware serials for reason above, but
what about a "factory/default" writable location for a string?
Reserve for example an location that can store like a MD5 hash hex
string or so and that is by default always the same (0000.. or the MD5
has for 'Ben Nanonote' or so).
Poeple who want then want to have a unique ID can update this at flash
time or so with their own ID system..
However not sure if it there is any advantage to "forsee" a thing like
this, I assume anyone who would need this will simply add it to their
own software stack somewhere.

Regards,
Gunther




More information about the discussion mailing list


interactive